Use of spiral and non-spiral computed tomography arterial portography in the detection of potentially malignant liver masses

The present study assesses the usefulness of computed tomography (CT) arterial portography (CTAP) in detecting and defining the number and anatomy of potentially malignant liver lesions. One hundred and one adults studied in 1993 and 1994 were retrospectively reviewed, including patients with primary or secondary tumours for possible resection and patients with non-hepatic malignancies in whom the detection of liver metastases would preclude surgery. Twenty-three patients underwent non-spiral CT studies and 78 had studies on a spiral unit, with 22 of these having single phase and 56 having dual phase studies to overcome artefact problems. The relationship between lesion size and detection sensitivity is critical. On non-spiral studies, the overall lesion detection sensitivity and positive predictive value was 69 and 90%, respectively. Detection sensitivity was 100 and 20% for lesions > 1 cm and < 1 cm, respectively. On single phase spiral CTAP the overall detection sensitivity and positive predictive value was 80 and 66%, respectively. Detection sensitivity for lesions > 1 cm and < 1 cm was 100 and 0%, respectively. On dual phase spiral CTAP the overall detection sensitivity and positive predictive value was 76 and 71%, respectively. For lesions > 1 cm and < 1 cm the sensitivity was 81 and 55%, respectively. Eighteen patients with non-hepatic malignancies with unsuspected metastatic spread did not proceed to major surgery because of liver metastases detected on CTAP. Perfusion artefacts occurred in 30 and 64% of non-spiral and of initial portal venous spiral CTAP studies, respectively. By using the double-phase technique, these artefacts were substantially diminished. In conclusion, CTAP is a valuable tool for assessing the presence, site and size of possible liver tumours and confers a benefit even when previous ultrasound and conventional CT have already been used. In addition, CTAP has a lower limit of useful resolution of approximately 1 cm. Perfusion artefacts can be reduced by a dual phase protocol.